EP2434538B1

EP2434538B1 - Semiconductor switch device and method for manufacturing semiconductor switch device

Info

Publication number: EP2434538B1
Application number: EP10777701.3A
Authority: EP
Inventors: Tsunekazu Saimei; Kazuya Kobayashi; Koshi Himeda; Nobuyoshi Okuda
Original assignee: Murata Manufacturing Co Ltd
Current assignee: Murata Manufacturing Co Ltd
Priority date: 2009-05-19
Filing date: 2010-05-14
Publication date: 2018-08-22
Anticipated expiration: 2030-05-14
Also published as: JP5652392B2; JPWO2010134468A1; CN102428550B; CN102428550A; WO2010134468A1; TW201108395A; US8933497B2; TWI509774B; EP2434538A4; EP2434538A1; US20120091513A1

Description

Technical Field

The present invention relates to a semiconductor switch device in which a switch circuit, etc. are constituted by using semiconductor elements, such as an FET (Field Effect Transistor), and to a method of manufacturing the semiconductor switch device.

Background Art

Transition from the second-generation cellular phone system to the third-generation cellular phone system is now in progress. With such system transition, an integrated circuit fabricated by integrating a logic circuit, an amplification circuit, etc. with the switch circuit is employed in front-end portions of the cellular phones in increasing number.
In that type of integrated circuit, it is demanded to improve not only characteristics of the switch circuit alone, but also characteristics of the entire integrated circuit, such as an insertion loss and isolation. Therefore, some integrated circuit is constituted as a semiconductor switch device in which a depletion-type FET (hereinafter referred to as a "D-type FET") and an enhancement-type FET (hereinafter referred to as an "E-type FET") are formed on a single semiconductor substrate in a mixed way (see, e.g., Patent Literature (PTL) 1). The D-type FET has a normally-on characteristic that a threshold voltage is negative when a drain current starts to flow, and it is featured in having a smaller insertion loss than the E-type FET. The D-type FET is used in many switch circuits. The E-type FET has a normally-off characteristic that a threshold voltage is positive when a drain current starts to flow, and it is used in many amplification circuits and many logic circuits.
PTL 2 discloses an RF switch with an on-chip decoder using enhancement-depletion mode MESFET or PHEMT semiconductor technology, for example GaAs process technology.
PTL 3 discloses a technology for manufacturing GaAs based E-type FETs and D-type FETs on the same substrate. The gate electrode of the E-type FET has a mushroom shape having an upper portion with a larger width than a lower portion.
PTL 4 discloses E-type and D-type PHEMT devices with a double recessed structure. In both device types, lower gate recesses are narrower than the corresponding upper gate recesses, which enables superior breakdown voltage.

Citation List

Patent Literature

PTL 1: Japanese Patent Application Publication No. 2005-203642
PTL 2: US Patent Application Publication No. 2005/0206439 A1
PTL 3: Japanese Patent Application No. H-6-216327
PTL 4: European Patent Application Publication No. 1261035 A2

Summary of Invention

Technical Problem

The third-generation cellular phone system faces a serious problem that intermodulation distortion also enters a reception path and causes a reception error, in addition to harmonic distortion (signal distortion) which has so far been a problem in the second-generation cellular phone system. The intermodulation distortion is generated by mixing of jamming waves and transmitted waves, which exist in air. Thus, in the third-generation cellular phone system, the distortion characteristic that has not been taken into consideration as a problem in the second-generation cellular phone system is an important characteristic, and an improvement of the distortion characteristic with reduction of both the harmonic distortion and the intermodulation distortion is demanded.
The inventors of this application have found that linearity of a capacitance characteristic in the FET constituting the switch circuit greatly affects the distortion characteristic, and have accomplished the present invention based on the finding.
An object of the present invention is to provide a semiconductor switch device having an improved distortion characteristic, and a method of manufacturing the semiconductor switch device.

Solution to Problem

The semiconductor switch device according to the present invention includes a plurality of semiconductor elements, e.g., an E-type FET and a D-type FET, formed on a single semiconductor substrate with provision of recesses for the semiconductor elements. The plurality of semiconductor elements are comprised in a switch circuit and a connection circuit, e.g., a logic circuit, connected to the switch circuit. Each of the semiconductor elements comprises a gate electrode forming portion, a drain electrode forming portion, and a source electrode forming portion including respectively a gate electrode, a drain electrode, and a source electrode. The gate electrode forming portion is arranged between the drain electrode forming portion and the source electrode forming portion. The switch circuit is constituted by the semiconductor element in which the gate electrode has a rectangular external shape in section. The connection circuit includes the semiconductor element in which the gate electrode has an external shape other than being rectangular in section, such as a V-shape or a T-shape in section.
In the arrangement described above, a stray capacitance component is reduced in the gate electrode having the rectangular sectional shape (hereinafter referred to as a "rectangular gate") as compared with a gate electrode having a V-shape or T-shape in section (hereinafter referred to as a "V-shaped or T-shaped gate"). The stray capacitance component remains after turning-off of a switch circuit and causes leakage of a high-frequency signal, thus degrading a distortion characteristic of the switch circuit. Further, the rectangular gate enables the recess to be formed in a larger width than the V-shaped gate and the T-shaped gate, for example. By forming the recess in a larger width, when the switch circuit is turned off, potential gradients between the gate electrode and the source electrode and between the gate electrode and the drain electrode can be moderated and linearity of the capacitance characteristic in the D-type FET can be improved. As a result, the distortion characteristic of the switch circuit can be improved.
There is a risk that the larger recess width may increase resistance in a channel region. In the third-generation cellular phone system to which the present invention is to be applied, however, it is more important to improve the distortion characteristic than to reduce the resistance in the channel region. In the present invention, therefore, the rectangular gate enabling the recess width to be more easily increased is employed in the switch circuit in which increasing the recess width is effective in improving the distortion characteristic. On the other hand, in the connection circuit in which the effect of an increase of the recess width upon the distortion characteristic is less important, the V-shaped gate or the T-shaped gate is formed to suppress an increase of resistance in the channel region of the E-type FET. Herein, the term "recess" implies a groove formed between the drain electrode forming portion and the source electrode forming portion and having a recessed sectional shape. The term "recess width" implies a width of the groove.
According to the invention, the recess has a multi-stepped shape including a first recess portion and a second recess portion deeper than the first recess portion, the second recess portion having a recess width smaller than a recess width of the first recess portion. With that arrangement, the stray capacitance component generated in the recess can be further reduced, and the linearity of the capacitance characteristic in the semiconductor element can be improved.
According to the invention, a ratio of the recess width of the second recess portion to the recess width of the first recess portion is larger in the semiconductor element including the rectangular gate than in the semiconductor element including the V-shaped gate and the T-shaped gate. With that feature, the increase of resistance in the channel region of the semiconductor element can be suppressed in the connection circuit while the distortion characteristic of the semiconductor element can be reliably improved in the switch circuit.
Preferably, the recess width of the second recess is larger in the semiconductor element including the gate electrode having the rectangular sectional shape than in the semiconductor element including the gate electrode having the sectional shape other than being rectangular. With that feature, the increase of resistance in the channel region of the semiconductor element can be suppressed in the connection circuit while the distortion characteristic of the semiconductor element in the switch circuit can be more reliably improved.
Preferably, an amplification circuit provided with the semiconductor element including the V-shaped gate or the T-shaped gate is further formed in the semiconductor substrate. With that feature, the amplification circuit can also be formed in an integrated state on the semiconductor substrate, whereby a degree of integration in the circuit configuration can be increased and some steps in a manufacturing process can be performed in common.
In a manufacturing method according to the present invention, after forming the V-shaped gate or the T-shaped gate, the rectangular gate is formed. The V-shaped gate and the T-shaped gate have relatively complex shapes and hence require a long and complicated manufacturing process. For that reason, if the rectangular gate is formed prior to forming the V-shaped gate and/or the T-shaped gate, a risk of damaging the rectangular gate due to, e.g., heat generated in the manufacturing process of the V-shaped gate and/or the T-shaped gate is increased. The possible damage can be suppressed by forming the rectangular gate, which is relatively simple in the manufacturing process, in a later stage.

Advantageous Effects of Invention

According to the present invention, the linearity of the capacitance characteristic in the semiconductor element can be improved while a decrease of an amplification rate and an increase of an impedance component are suppressed. It is hence possible to improve the distortion characteristic and to suppress, e.g., the occurrence of a reception error in the third-generation cellular phone system.

Brief Description of Drawings

[Fig. 1] Fig. 1 is a schematic sectional view of a semiconductor switch device according to a first embodiment of the present invention.
[Fig. 2] Fig. 2 represents characteristic graphs of the semiconductor switch device illustrated in Fig. 1.
[Fig. 3] Fig. 3 represents schematic circuit diagrams of the semiconductor switch device illustrated in Fig. 1.
[Fig. 4] Fig. 4 represents sectional views illustrating states in successive steps of a manufacturing process for the semiconductor switch device illustrated in Fig. 1.
[Fig. 5] Fig. 5 is a schematic sectional view of a semiconductor switch device according to a second embodiment of the present invention.
[Fig. 6] Fig. 6 represents schematic circuit diagrams of the semiconductor switch device illustrated in Fig. 5.
[Fig. 7] Fig. 7 is a schematic sectional view of a semiconductor switch device according to a third embodiment of the present invention.
[Fig. 8] Fig. 8 is a schematic sectional view of a semiconductor switch device according to an example not forming part of the claimed invention.

Description of Embodiments and an Example

<<First Embodiment>>

A semiconductor switch device 1 according to a first embodiment of the present invention will be described below in connection with an example in which an FET is formed as a semiconductor element. It is to be noted that the present invention can also be preferably applied to the case using an HEMT (high Electron Mobility Transistor) as one type of FET.
Fig. 1 is a schematic sectional view of the semiconductor switch device 1.
The semiconductor switch device 1 includes a plurality of semiconductor elements including at least two types of semiconductor elements E1 and D1. Fig. 1 illustrates an exemplary construction in which the semiconductor element E1 and the semiconductor element D1 are formed side by side.
The semiconductor switch device 1 includes a semiconductor substrate 2, gate electrodes 4A and 4B, source electrodes 5A and 5B, and drain electrodes 6A and 6B. The semiconductor substrate 2 includes a GaAs layer 2A as a semiconductor layer, a channel layer 2B epitaxially grown on the GaAs layer 2A, and a contact layer 2C epitaxially grown on the channel layer 2B.
The semiconductor substrate 2 includes a groove 3C that is formed by partly removing the contact layer 2C, the channel layer 2B, and the GaAs layer 2A. The groove 3C delimits a region where one semiconductor element is formed, and it makes the GaAs layer 2A exposed to the outside.
The semiconductor substrate 2 includes recesses 3A and 3B, which are formed by partly removing the contact layer 2C, in the regions where the semiconductor elements are formed, respectively. The recesses 3A and 3B make the channel layer 2B exposed to the outside.
The source electrodes 5A and 5B and the drain electrodes 6A and 6B are formed on the contact layer 2C at positions corresponding to respective ridges aside on both sides of the recesses 3A and 3B, respectively. The source electrodes 5A and 5B and respective portions of the contact layer 2C just under the formers constitute source electrode forming portions in the present invention. The drain electrodes 6A and 6B and respective portions of the channel layer 2B just under the formers constitute drain electrode forming portions in the present invention.
The gate electrodes 4A and 4B are formed on bottom surfaces of the recesses 3A and 3B. The gate electrode 4A is formed in a state partly buried in the channel layer 2B, and the gate electrode 4B is formed on the channel layer 2B. Each of portions of the gate electrodes 4A and 4B, which project respectively from the bottom surface of the recesses 3A and 3B, constitutes a gate electrode forming portion in the present invention.
The semiconductor element E1 is an E-type FET and is made up of the semiconductor substrate 2, the gate electrode 4A, the source electrode 5A, and the drain electrode 6A. The gate electrode 4A is a V-shaped gate formed to have a V-shape in section (hereinafter referred to as a "V-shaped gate 4A"). The recess 3A is formed in the region of the semiconductor substrate 2 where the semiconductor element E1 is formed. The recess 3A has two steps in sectional shape, which are constituted by a first recess portion formed by processing the contact layer 2C and a second recess portion formed by processing the channel layer 2B. A recess width L1 of the first recess portion is larger than a recess width L2 of the second recess portion.
The semiconductor element D1 is a D-type FET and is made up of the semiconductor substrate 2, the gate electrode 4B, the source electrode 5B, and the drain electrode 6B. The gate electrode 4B is a rectangular gate formed to have a rectangular shape in section (hereinafter referred to as a "rectangular gate 4B"). The recess 3B is formed in the region of the semiconductor substrate 2 where the semiconductor element D1 is formed. The recess 3B has two steps in sectional shape, which are constituted by a first recess portion formed by processing the contact layer 2C and a second recess portion formed by processing the channel layer 2B. A recess width L1' of the first recess portion is larger than a recess width L2' of the second recess portion.
In the semiconductor element D1 of this embodiment, because the rectangular gate 4B is employed, its surface area can be reduced. Comparing with the case employing a V-shaped gate or a T-shaped gate, therefore, a stray capacitance component generated between the semiconductor substrate 2 and each of the source electrode 5B and the drain electrode 6B can be reduced. Further, the recess width L2' for the semiconductor element D1 is set larger than the recess width L2 for the semiconductor element E1 such that a potential gradient in the channel layer 2B is moderated and linearity of a capacitance characteristic is improved. On the other hand, in the semiconductor element E1, a decrease of an amplification rate and an increase of an impedance component are suppressed by employing the V-shaped gate.
The capacitance characteristic of a semiconductor element is now described in connection with, by way of example, the D-type FET.
Fig. 2(A) is a graph plotting the relationship between a source-drain capacitance Coff and a gate-source voltage Vgs in an off-state of the D-type FET. The graph comparatively plots the case employing a rectangular gate as the D-type FET and the case employing a V-shaped gate as the D-type FET. Further, the gate-source voltage Vgs is represented in terms of the so-called inverse voltage.
As seen from the graph of Fig. 2(A), the capacitance Coff in the case employing the rectangular gate is always smaller than the capacitance Coff in the case employing the V-shaped gate, and the stray capacitance component between the gate electrode and each of the drain electrode and the source electrode can be suppressed.
Also, from the graph of Fig. 2(A), it can be confirmed that the rectangular gate provides a smaller slope of change in the capacitance Coff than the V-shaped gate in a region where the voltage Vgs is larger than a pinch-off voltage of about 0.8 V. Thus, it can be confirmed that, by employing the rectangular gate and increasing the recess width of the first recess portion, bias dependency of the capacitance Coff can be moderated and the linearity can be improved.
Fig. 2(B) is a graph plotting the relationship between the source-drain capacitance Coff and L2'/L1', i.e., a ratio of the recess width of the second recess portion to the recess width of the first recess portion in the two-stepped recess for the rectangular gate. Fig. 2(B) comparatively represents data while the gate-source voltage Vgs is kept at the same condition.
As seen from the graph of Fig. 2(B), the capacitance Coff decreases as the recess width ratio L2'/L1' increases. Thus, it is understood that the capacitance Coff can be reduced by increasing the recess width of the second recess portion.
While the above description has been made on data obtained at different recess width ratios in the rectangular gate, the confirmed relationship is held regardless of the gate shape. It is therefore preferable to set the recess width ratio in a semiconductor element in which the capacitance Coff is to be reduced, such as a semiconductor element constituting a switch circuit, to be larger than the recess width ratio in another semiconductor element in which the necessity of reducing the capacitance Coff is relatively low.
One example of circuit configuration of the semiconductor switch device 1 will be described below.
Fig. 3(A) is a schematic circuit diagram to explain an exemplary configuration of the semiconductor switch device 1. The semiconductor switch device 1 includes a switch circuit SW and a logic circuit LOGIC.
Fig. 3(B) is a schematic circuit diagram to explain an exemplary configuration of the switch circuit SW. The switch circuit SW is constituted by a plurality of semiconductor elements D1, and it has input/output ports PORT1 and PORT2 and an antenna port ANT. In the switch circuit SW, each semiconductor element D1 is turned on or off in accordance with a control voltage input to a control terminal, whereby connection of the input/output port PORT1 or PORT2 to the antenna port ANT is selected.
Here, it is supposed that all semiconductor elements constituting the switch circuit SW are the semiconductor elements D1 including the rectangular gates 4B. Thus, the linearity is improved in the capacitance characteristic of each semiconductor element D1, and the switch circuit SW has a very good distortion characteristic.
Fig. 3(C) is a schematic circuit diagram to explain an exemplary configuration of the logic circuit LOGIC. The logic circuit LOGIC is constituted by the semiconductor element D1 and the semiconductor element E1. The logic circuit LOGIC outputs a voltage of a logic level to the control terminal of the switch circuit SW in accordance with a control voltage Vctl that is input to an input port the logic circuit LOGIC.
In the illustrated logic circuit LOGIC, since the semiconductor element E1 including the V-shaped gate is employed, the decrease of the amplification rate and the increase of the impedance component can be suppressed in amounts corresponding to the use of the semiconductor element E1 in comparison with the case where all the gate electrode forming portions of the semiconductor elements E1 are formed to be rectangular in section.
One example of a process for manufacturing the semiconductor switch device 1 will be described below.
Fig. 4(A) is a sectional view illustrating a state during a region dividing step in the manufacturing process.
In this step, the groove 3C is formed at each of positions partitioning a plurality of semiconductor elements in the semiconductor substrate 2. More specifically, the semiconductor substrate 2 having a flat plate shape and including the GaAs layer 2A, the channel layer 2B, and the contact layer 2C is first prepared. Then, the groove 3C is formed by etching, for example, with a depth extending from the contact layer 2C up to the GaAs layer 2A. After completion of the region dividing step, the manufacturing process advances to a next ohmic electrode forming step.
Fig. 4(B) is a sectional view illustrating a state during the ohmic electrode forming step in the manufacturing process.
In this step, ohmic electrodes serving as the drain electrodes 6A and 6B and the source electrode 5A and 5B are formed in the regions individually delimited by the groove 3C. The ohmic electrodes are each formed by metal vapor deposition. After completion of the ohmic electrode forming step, the manufacturing process advances to a next common etching step.
Fig. 4(C) is a sectional view illustrating a state during the common etching step in the manufacturing process.
In this step, respective first recess portions 13A and 13B of the recesses 3A and 3B are formed. More specifically, a resist film is first formed by photolithography. Next, the contact layer 2C is partly removed by wet etching or dry etching. Thereafter, the resist film is removed. After completion of the common etching step, the manufacturing process advances to a next E-type FET etching step.
Fig. 4(D) is a sectional view illustrating a state during the E-type FET etching step in the manufacturing process.
In this step, the second recess portion 13C of the recess 3A is formed. More specifically, a resist film 11A is first formed on the semiconductor substrate 2 by photolithography. A resist window having a taper in match with the shape of a lower surface of the V-shaped gate 4A is formed in the resist film 11A. Further, a resist film 11B is formed over the resist film 11A by photolithography. A resist window having an opening in match with the shape of the V-shaped gate 4A, as viewed from above, is formed in the resist film 11B. Then, the channel layer 2B is partly removed by, e.g., wet etching or dry etching. After completion of the E-type FET etching step, the manufacturing process advances to a next E-type FET gate electrode forming step.
Fig. 4(E) is a sectional view illustrating a state during the E-type FET gate electrode forming step in the manufacturing process.
In this step, the V-shaped gate 4A is formed. More specifically, metal vapor deposition is first carried out by utilizing the resist films 11A and 11B, which have been formed in the preceding step. Thereafter, the resist films 11A and 11B are removed. Thus, a process of forming the resist films is curtailed by utilizing the resist films, which have been employed in the preceding step, to form the V-shaped gate 4A in this E-type FET gate electrode forming step as well. After completion of the E-type FET gate electrode forming step, the manufacturing process advances to a next D-type FET etching step.
Fig. 4(F) is a sectional view illustrating a state during the D-type FET etching step in the manufacturing process.
In this step, the second recess portion 13D of the recess 3B is formed. More specifically, a resist film 11C is first formed on the semiconductor substrate 2 by photolithography. A resist window having an opening in match with the shape of the rectangular gate 4B, as viewed from above, is formed in the resist film 11C. Then, the channel layer 2B is partly removed by, e.g., wet etching or dry etching. After completion of the D-type FET etching step, the manufacturing process advances to a next D-type FET gate electrode forming step.
Fig. 4(G) is a sectional view illustrating a state during the D-type FET gate electrode forming step in the manufacturing process.
In this step, the rectangular gate 4B is formed. More specifically, metal vapor deposition is first carried out by utilizing the resist film 11C, which has been formed in the preceding step. Thereafter, the resist film 11C is removed. Thus, a process of forming the resist film is curtailed by utilizing the resist film, which has been employed in the preceding step, to form the rectangular gate 4B in this D-type FET gate electrode forming step as well.
The semiconductor switch device 1 is manufactured by the above-described manufacturing process. With this embodiment, since the rectangular gate 4B is formed after forming the V-shaped gate 4A that requires a relatively long manufacturing process, an influence of the steps of forming the semiconductor element in a later stage upon the semiconductor element formed in an earlier stage can be suppressed even when the different types of semiconductor elements are formed in sequence.

<<Second Embodiment>>

A semiconductor switch device 21 according to a second embodiment of the present invention will be described below. In the following, the same components as those in the first embodiment are denoted by the same symbols, and description of those components is omitted.
Fig. 5 is a schematic sectional view of the semiconductor switch device 21.
The semiconductor switch device 21 includes a plurality of semiconductor elements including at least three types of semiconductor elements E1, D1 and D2.
The semiconductor element D2 is a D-type FET and is made up of a semiconductor substrate 22, a gate electrode 24, a source electrode 25, and a drain electrode 26. The gate electrode 24 is a V-shaped gate formed to have a V-shape in section (hereinafter referred to as a "V-shaped gate 24"). By partly removing the contact layer 2C, a recess 23 is formed in a region of the semiconductor substrate 22 where the semiconductor element D2 is formed. The recess 23 has two steps in sectional shape and has a recess width in the same size as that for the semiconductor element E1. The source electrode 25 and the drain electrode 26 are formed on the contact layer on both sides of the recess 23.
Since the V-shaped gate 24 is employed in the semiconductor element D2 of this embodiment, the recess width L2 is reduced in comparison with the case employing the rectangular gate. As a result, the decrease of the amplification rate and the increase of the impedance component can be suppressed in the semiconductor element D2.
One example of circuit configuration of the semiconductor switch device 1 will be described below.
Fig. 6(A) is a schematic circuit diagram to explain an exemplary configuration of the semiconductor switch device 1. The semiconductor switch device 1 includes a switch circuit SW, a logic circuit LOGIC, a power amplifier PA, and a low-noise amplifier LNA.
Fig. 6(B) is a schematic circuit diagram to explain an exemplary configuration of the switch circuit SW. The switch circuit SW is constituted by a plurality of semiconductor elements D1.
Here, it is supposed that all semiconductor elements constituting the switch circuit SW are the semiconductor elements D1 including the rectangular gates 4B. Thus, the linearity is improved in the capacitance characteristic of each semiconductor element D1, and the switch circuit SW has a very good distortion characteristic.
Fig. 6(C) is a schematic circuit diagram to explain an exemplary configuration of the logic circuit LOGIC. The logic circuit LOGIC is constituted by the semiconductor element D2 and the semiconductor element E1. The logic circuit LOGIC outputs a voltage of a logic level to a control terminal of the switch circuit SW in accordance with a control voltage Vctl that is input to an input port of the logic circuit LOGIC.
In this embodiment, since the logic circuit LOGIC is constituted by the semiconductor elements E1 and D2 each including the V-shaped gate, the decrease of the amplification rate and the increase of the impedance component can be suppressed corresponding to the use of the semiconductor elements E1 and D2.
Fig. 6(D) is a schematic circuit diagram to explain an exemplary configuration of the power amplifier PA and the low-noise amplifier LNA. Each of the power amplifier PA and the low-noise amplifier LNA is constituted by the semiconductor element D2. Accordingly, the decrease of the amplification rate and the increase of the impedance component can be suppressed corresponding to the use of the semiconductor element D2.

<<Third Embodiment>>

A semiconductor switch device 31 according to a third embodiment of the present invention will be described below. In the following, the same components as those in the first and second embodiments are denoted by the same symbols, and description of those components is omitted.
Fig. 7 is a schematic sectional view of the semiconductor switch device 31.
The semiconductor switch device 31 includes a plurality of semiconductor elements including at least three types of semiconductor elements E2, D1 and D3.
The semiconductor element D3 is a D-type FET, and it includes a gate electrode 34A. The semiconductor element E2 is an E-type FET, and it includes a gate electrode 34B. The gate electrodes 34A and 34B are each a T-shaped gate formed to have a T-shape in section.
Even when the T-shaped gate 24 is employed instead of the V-shaped gate as in this embodiment, the decrease of the amplification rate and the increase of the impedance component in the semiconductor element can be suppressed by minimizing the recess width for the T-shaped gate as in the case employing the V-shaped gate.

<< Example >>

A semiconductor switch device 41 according to an example not forming part of the claimed invention will be described below. In the following, the same components as those in the first to third embodiments are denoted by the same symbols, and description of those components is omitted.
Fig. 8 is a schematic sectional view of the semiconductor switch device 41.
The semiconductor switch device 41 includes a plurality of semiconductor elements including at least three types of semiconductor elements E2, D4 and D3.
The semiconductor element D4 is a D-type FET including a rectangular gate, and it includes a semiconductor substrate 42 in which a recess 43 is formed. The recess 43 is formed such that its recess width has the same size as that for each of the semiconductor element D3 and the semiconductor element E2. Such a structure of the semiconductor element D4 is employed in the semiconductor element constituting the switch circuit SW.
Even when the recess width is set to be the same for all the semiconductor elements as in this example, the distortion characteristic of the switch circuit can be improved by employing the T-shaped gate, the V-shaped gate, and the rectangular gate in a combined manner.

Reference Signs List

1, 21, 31, 41 ... semiconductor switch device 2, 22, 42 ... semiconductor substrate
3A, 3B, 23, 43 ... recess
3C ... groove 4A, 24 ... gate electrode (V-shaped gate) 34A, 34B - gate electrode (T-shaped gate) 4B ... gate electrode (rectangular gate)
5A, 5B, 25 ... source electrode
6A, 6B, 26 ... drain electrode E1, E2, D1, D2, D3, D4 ... semiconductor element
LOGIC ... logic circuit
SW ... switch circuit

Claims

A semiconductor switch device (1) including a plurality of semiconductor elements (E1, E2, D1) formed on a semiconductor substrate (2) with provision of recesses (3A, 3B) for the semiconductor elements (E1, E2, D1), the plurality of semiconductor elements (E1, E2, D1) being comprised in an antenna switch circuit (SW) and a connection circuit connected to the antenna switch circuit (SW), each of the semiconductor elements (E1, E2, D1) comprising: a source electrode forming portion including a source electrode (5A, 5B); a drain electrode forming portion including a drain electrode (6A, 6B); and a gate electrode forming portion including a gate electrode (4A, 34B, 4B) projecting from a bottom surface of the recess (3A, 3B), and arranged between the drain electrode forming portion and the source electrode forming portion, wherein the antenna switch circuit (SW) is constituted by the semiconductor element (D1) in which the gate electrode (4B) has a rectangular external shape in section, and the connection circuit includes the semiconductor element (E1, E2) in which the gate electrode (4A, 34B) has an external shape other than being rectangular in section, wherein the recess (3A, 3B) has a multi-stepped shape including a first recess portion formed between the drain electrode forming portion and the source electrode forming portion, and a second recess portion formed around the gate electrode forming portion to be deeper than the first recess portion, the second recess portion having a recess width (L2', L2) smaller than a recess width (L1', L1) of the first recess portion, wherein a ratio of the recess width (L2', L2) of the second recess portion to the recess width (L1', L1) of the first recess portion is larger in the semiconductor element (D1) including the gate electrode (4B) having a rectangular sectional shape than in the semiconductor element (E1, E2) including the gate electrode (4A, 34B) having a sectional shape other than being rectangular.
The semiconductor switch device (1) according to Claim 1, wherein the recess width (L2', L2) of the second recess (3A, 3B) is larger in the semiconductor element (D1) including the gate electrode (4B) having the rectangular sectional shape than in the semiconductor element (E1) including the gate electrode (4A, 34B) having the sectional shape other than being rectangular.
The semiconductor switch device (1) according to any one of Claims 1 to 2, wherein the semiconductor element (D1) constituting the antenna switch circuit (SW) is a depletion-type FET.
The semiconductor switch device (1) according to any one of Claims 1 to 3, wherein an amplification circuit provided with the semiconductor element (E1, E2) including the gate electrode (4A, 34B) having the sectional shape other than being rectangular is further formed in the semiconductor substrate (2).
A method of manufacturing the semiconductor switch device (1) according to any one of Claims 1 to 4, the method comprising the step of: after forming the gate electrode (4A, 34B) having the sectional shape other than being rectangular, forming the gate electrode (4B) having the rectangular sectional shape.